Scientific inquiry: from questions to evidence
FoundationScience is not guessing: ask a testable question, control variables, collect data and use evidence to explain the result.
- Questions should be testable
- Change one independent variable at a time
- Repeat tests and use data rather than intuition
§Text Passage
1Scientific knowledge does not appear out of thin air; it is built step by step through careful inquiry. An investigation begins with observation and questions: from what we see, we ask a testable question, propose a hypothesis based on what we already know, design an experiment to test it, and finally draw a conclusion from the data collected.
2A fair test lies at the heart of the whole process. Only one variable may be deliberately changed while every other condition stays constant. For example, when investigating how fertiliser affects plant growth, sunlight, watering and pot size must all match, and a control group without fertiliser is needed before the results carry any weight. Repeating the trial and averaging the results reduces random error.
3The same method works in daily life. To find out which washing-up liquid cuts grease best, compare them with the same amount of oil, the same water temperature and washing time, changing only the brand; record the measurements honestly, and when odd data appear, never delete them — test again and find out why.
4Above all, conclusions must rest on evidence: they stand only when the data support the hypothesis, and must be revised or abandoned when the data say otherwise. This method grounds scientific conclusions in evidence rather than guesswork or authority, helps us tell correlation from causation, and lets us judge truth wisely in an age of information overload.
1Summary 重點摘要
Science is not guessing: ask a testable question, control variables, collect data and use evidence to explain the result.
2Vocabulary cards Vocabulary
Recording what happens systematically with senses or instruments.
A testable prediction based on what is already known.
A factor that is changed, measured or kept constant in an investigation.
Data from observations or measurements used to support or challenge an explanation.
An investigation that changes only one variable while keeping all other conditions the same.
The factor deliberately changed by the investigator in an experiment.
The result that is measured as the independent variable changes.
A factor kept constant throughout the experiment to keep the test fair.
3Step by step Steps
- Ask a testable questionScientific inquiry starts from a testable question: it must be answerable by a fair test, such as 'Does the strength of light affect the height of bean seedlings?' A question as broad as 'Why do plants grow?' cannot be settled in a single experiment. So before doing anything, narrow the scope to one measurable factor — only then does the investigation have a clear direction. To judge whether a question is testable, ask: can it be measured, and can it be compared? Only questions that pass both checks make a good starting point.
- Separate the variablesThe key to a fair test is telling the three kinds of variable apart: the independent variable is the single condition you deliberately change, such as light level; the dependent variable is the result you measure, such as seedling height; and all the control variables — water, soil, temperature, seed batch and time — must stay exactly the same. If even one of them differs, the comparison becomes unfair and the measurements cannot be trusted. Before the experiment, list all three kinds of variable on a sheet beside you and check it after every change — a simple checklist guards against omissions.
- Repeat, record and compareSolid data come from repetition and careful recording: organise the results in a table, repeat each trial or grow several samples, then compare averages and look for trends. A single result is easily skewed by chance and cannot support a reliable conclusion. When odd data appear, never delete them — record them honestly and test again, because an anomaly sometimes hints at a new discovery. Record the day's conditions, such as temperature and weather, alongside the data; these notes often explain anomalies when you review the results later.
- Let the evidence decideBefore concluding, let the evidence speak: a conclusion stands only when the data support the hypothesis, and it must be revised or abandoned when they do not. Also keep correlation apart from causation: ice-cream sales and heatstroke cases rising together do not mean ice-cream causes heatstroke — hot weather likely drives both. With this in mind, you are far less easily misled by figures in news reports and advertisements. Practise asking one more question: what else could explain this result? Hunting for alternative explanations makes a conclusion far more robust.
- Propose a hypothesis firstA hypothesis is a tentative answer to the question, based on what you already know, and it must be specific enough to test — for example, 'the stronger the light, the taller the bean seedlings grow'. When writing a hypothesis, state the expected link between the independent and dependent variables. Whether it stands can only be judged after the experiment is done and the data are in, so the hypothesis guides the design of the experiment and is one of the starting points of the whole inquiry. A good hypothesis tells you how to test it at a glance; if no experiment comes to mind, the hypothesis is still too vague and needs narrowing.
- Set up a control groupA control group is the untreated standard group against which the experimental group is compared: when testing fertiliser, the control plants receive none while every other condition matches the experimental group exactly. Only with such a comparison can you confirm that the difference in growth truly came from the fertiliser and not from chance or some other factor. The same design logic is widely used in medical trials and product testing, making the control group a vital part of the scientific method. A control group also exposes surprises: if the control plants change too, something else is at work and the design must be re-examined.
4Diagrams Diagrams
5Worked example Example
Example: how could we test whether fertiliser makes plants taller?
- Independent variable: fertiliser or no fertiliser
- Dependent variable: plant height after two weeks
- Controls: same species, soil, water, light, pot and time
- Include a control group without fertiliser for comparison
- Answer: Change only the fertiliser (independent variable) and measure each plant's height after two weeks (dependent variable); keep species, soil, water, light, pots and time identical, include a no-fertiliser control group, and average several plants to cut random error — avoid over-generalising, let the evidence, not a single result, carry the conclusion, and report any anomalous data honestly.
6Common mistakes Mistakes
7Quick check Quick check
Q1What is the factor deliberately changed in an experiment called?
Q2Why should control variables be kept the same?
Q3How are data and evidence related?
Q4In the hypothesis 'sugar dissolves faster in hotter water', what is the dependent variable?
Q5What role does a control group play in an experiment?
8Recap Recap
- Questions should be testable
- Change one independent variable at a time
- Repeat tests and use data rather than intuition
- In a fair test, all control variables stay the same
- Conclusions must rest on evidence and stay open to revision